If I Were Building My Business Again in the Age of AI

If I were building Icon Visual Marketing again today, knowing what I now know about AI, I wouldn't simply introduce new technology into the business I used to run.

I'd build the business differently.

We built a successful strategy-led marketing business with talented people, long-standing client relationships and specialist capabilities across strategy, account management, creative and production.

I'm proud of what we built. But looking back, I can see things I would approach differently today.

Not because we lacked good people or good systems. It was what happened between them that often created the greatest challenges.

A conversation that wasn't fully understood. A brief that missed something important. A decision made without knowing what had already been discussed. A promise interpreted differently by the client and the team.

Sometimes the work was excellent, yet the client was disappointed.

And sometimes we lost the relationship.

I wonder how much of that could have been avoided if we'd been better at connecting what our people knew.

More importantly, how many opportunities did we miss because the right people never had the full picture?

That's where I would begin if I were building again with AI.

The business we built made sense at the time

Running a strategy-led agency required a considerable range of skills.

We needed people who could understand markets, develop strategies, manage client relationships, produce creative work and deliver projects.

As the business grew, we developed the departments, processes and systems needed to coordinate that work.

But there's a consequence to organising people into specialist roles.

Information moves through handovers.

A client speaks to an account manager. The account manager prepares a brief. The strategist interprets it. The creative team develops the response. Production delivers the work.

Each person may do their job well, but the original conversation has passed through several interpretations.

The brief captures what someone understood. It doesn't necessarily capture everything that was said, or the history and expectations behind it.

I saw this repeatedly.

A client would express something in a meeting that didn't make its way into the brief. Someone would make an assumption that seemed reasonable but wasn't quite right. A concern would be mentioned, then forgotten until it became a problem.

We had talented people doing good work, yet too much effort was spent correcting misunderstandings.

And when the client became frustrated, they didn't care which department was responsible.

They expected us to understand their business and deliver what we'd promised.

That was reasonable.

I'd rethink what we employed people to do

If I were building again, I'd still want excellent strategists, designers and technical specialists.

But I'd place even greater emphasis on their ability to understand customers, build relationships and exercise judgement.

I'd want people who were curious about the client's business.

People who could recognise when the brief wasn't addressing the real problem.

People who could challenge an assumption without making the client feel challenged.

And people who understood that their responsibility didn't end when their particular part of the work was completed.

Those qualities were always valuable. AI makes me think they deserve even greater attention.

Because if technology can increasingly assist with research, analysis, design and production, I would want my people concentrating more of their effort on understanding what customers need and creating something worthwhile.

That doesn't make specialist expertise less important. It changes how I'd want that expertise applied.

I wouldn't begin by asking how many people AI could replace.

I'd ask what my people could become capable of doing if technology supported more of the work around them.

I'd want AI involved in the whole relationship

Today, most businesses seem to be introducing AI into individual functions.

Marketing uses it to produce content. Sales uses it to prepare proposals. Finance uses it for analysis. Operations uses it to document processes.

There's value in all of that.

But if each department adopts AI independently, we may simply end up with more efficient departments that still don't understand one another.

That's not the business I would want to build.

I'd want AI helping us connect the entire customer relationship.

The first conversation. The analysis. The brief. The decisions. The work. The results. And what we learned afterwards.

Imagine a client expressing concern about something during an early meeting.

That concern is captured, but it doesn't disappear into a transcript nobody reads. It remains available when the team prepares the brief, makes decisions and reviews the work.

If the brief doesn't address the concern, someone can question it.

If a new instruction conflicts with an earlier agreement, the team can clarify it before proceeding.

If the people delivering the work discover something important about the customer, that knowledge can inform the next commercial conversation.

I would want AI helping people recognise those connections and identify what needs clarification.

Of course, AI won't always interpret the information correctly. It may miss context or suggest connections that aren't supported.

That's why I would build human review and correction into the process.

The objective wouldn't be to have AI make the decisions. It would be to give people a better understanding from which to make them.

I saw what happens when understanding begins to move

One of the experiences that shaped this thinking came while I was working with an established business whose founder was increasingly overwhelmed.

She carried an enormous amount of knowledge about the clients, the relationships, the history and the decisions behind the work.

Her team was capable, but so much of the context remained with her that people naturally returned whenever something needed clarification.

We began introducing better ways to capture client intelligence earlier and make it available to the people who needed it.

Initially, the benefits appeared fairly practical. Better information, clearer briefs and fewer opportunities for misunderstanding.

But something happened that I hadn't fully appreciated at the beginning.

The meetings changed.

People were more engaged. They asked questions and challenged assumptions. They could contribute with greater confidence because they understood more of the client's situation and the thinking behind the work.

They weren't waiting for the founder to interpret everything.

She began to gain some headspace. The team became more empowered, and the culture started to shift.

What struck me was that these weren't suddenly different people.

They were people who now had more of the understanding they needed to contribute.

And that changed the way they participated.

It reinforced something important for me: information doesn't automatically create capability. People also need trust, encouragement and the authority to act.

But when those things come together, the effect can be significant.

I would want that principle at the centre of any business I built today.

I'd change what we sold

This is perhaps the biggest commercial difference.

If I were starting Icon Visual again, I wouldn't want the business's value to depend primarily on producing marketing materials, campaigns or creative deliverables.

Those things would still matter, but AI is changing the economics of producing them.

I'd want clients to engage us because we understood their business and could help them create commercial value.

That means understanding their customers, challenging the original brief when necessary, connecting marketing with sales and delivery, and helping them recognise opportunities they might otherwise miss.

Creative work would be part of the solution, not necessarily the definition of the service.

I'd also reconsider how we charged for that work.

Traditional agency economics often revolve around time, people, production and project delivery.

Those measures still matter when running a business.

But they don't always reflect what the customer actually values.

A client doesn't necessarily care how many hours it took to recognise an opportunity that could materially improve their business.

They care whether the opportunity is real and whether someone can help them act on it.

That doesn't mean every engagement should be priced on outcomes. Commercial responsibility, scope and risk still need to be agreed.

But I would want the conversation about fees to begin with the value we're helping create, rather than simply the resources required to produce the work.

I'd build a business that remembers and learns

One thing I would pay much more attention to is what happens after a project finishes.

In a busy service business, the team moves on. New briefs arrive. New clients need attention.

But what happens to everything we learned?

What did the customer value?

Which assumptions were wrong?

What created unnecessary rework?

Which relationships became stronger?

What did the delivery team discover that could help us serve the customer better next time?

Too often, those lessons remain with individuals.

Some make their way into future decisions. Others disappear when people change roles or leave the business.

I would want AI helping the organisation retain and connect that knowledge, with appropriate permissions and safeguards.

Not as an archive of everything everyone has ever said.

As a way of making relevant experience useful when the next decision needs to be made.

The real test wouldn't be how much information we'd captured.

It would be whether our people were making better decisions because of it.

Would it be a smaller business?

Possibly.

I suspect the mix of roles would be different, and some production activities would require fewer resources.

But I wouldn't set out to build a smaller business.

I'd set out to build a more valuable one.

I'd want to understand what customers would still pay for, what new services we could provide and how our people could contribute in ways that weren't previously practical.

I'd measure productivity, certainly.

But I'd also pay close attention to customer retention, the quality of relationships, rework, opportunities identified, the strength of the team and how dependent the business remained on particular individuals.

Those things have commercial consequences, even if they aren't always as easy to measure as billable hours.

And I would want people to see AI as something that helped them become more capable, rather than simply a technology introduced to make them work faster.

Would you build your business the same way today?

Most established businesses were designed around the capabilities, technology and commercial conditions available when they were built.

That doesn't mean their structures are wrong.

But AI gives us a reason to examine whether those structures still make sense.

If you were starting your business today, would you employ the same mix of people?

Would you organise them into the same departments?

Would you sell the same products and services?

Would you charge the same way?

And would you still allow so much customer knowledge and organisational experience to remain disconnected?

I know I wouldn't build Icon Visual exactly as I did before.

I'd build around people who understood customers, developed relationships, exercised judgement and took responsibility for creating value.

I'd use AI to support the connections between conversations, decisions, delivery and learning.

And I'd want the business to become more capable with every customer relationship, rather than simply more efficient at completing the next project.

If AI is changing the economics of the work we produce, shouldn't we also reconsider the economics of the businesses we've built to produce it?

That's the question I'd be asking if I were starting again.

And I think it's a question worth asking even if you've been in business for twenty years.

Joe Papadatos is the founder of Two Icons, helping established businesses discover commercial opportunities and build the capability to realise them.

Grow Better.